On Thu, Jan 8, 2009 at 10:34 AM, Davide Alberani <davide.alber...@gmail.com> wrote: > As a side-note, are there SQLite experts around? > Are such poor performances to be expected? A "you can't create a > 3GB single-file database and expect it to be fast, you insensitive clod" > would be clear enough. :-)
http://www.mail-archive.com/sqlite-us...@sqlite.org/msg23470.html While researching the long load times, I came across that particular thread. I suspect it's what's going on here. I don't think there's a way of improving it by presorting, since, well, there are multiple keys to sort on. Actually, I imagine that the source material is sorted anyway, right? And so you're going to have at least one index presorted by default anyway. I wonder though. It looks like sqlite supports tweaking connections via PRAGMA directives: http://www.sqlite.org/pragma.html#pragma_page_size Maybe during creation, would could issue something like: PRAGMA synchronous = OFF; If one has enough ram, a ramdisk/tmpfs of some sort might be helpful (I don't have enough swap configured to test at the moment). On Linux, I might try mounting a scratch partition with async, but I don't yet know if that means ``I know what I'm doing, ignore what the application tells you to do.'' mrc ------------------------------------------------------------------------------ Check out the new SourceForge.net Marketplace. It is the best place to buy or sell services for just about anything Open Source. http://p.sf.net/sfu/Xq1LFB _______________________________________________ Imdbpy-help mailing list Imdbpy-help@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/imdbpy-help
